a) There are currently few people dedicated to research on vulnerability and risk to high impact weather that are linked to the weather forecasting community. We need to build capacity to define, conduct, review and communicate vulnerability and risk research and applications.  

b) Much of the current in-depth work on human impacts, exposure, vulnerability and risk from high impact weather is based on particular users and locations. We need to assess, develop and improve approaches to assess weather-related vulnerability and risk and the additional benefit of providing such information in addition to the hazard forecast. This will involve examining the full range of approaches from simple overlays of physical hazard and exposure variables (e.g., to determine population within weather warning areas) to sophisticated real-time dynamic modelling of impacts (e.g., streamflow, hourly energy demand). It should cover:

—— treatment of chronic, long-term impacts separated in time and/or space from the specific event (e.g., effects of time required to recover and rebuild from loss of dwelling can induce stress, mental illness; creates vulnerabilities to subsequent events) 

—— methods to distinguish evolving high consequence events/thresholds for which special or more focused/relevant communications (pre-, during, post-event) must be provided to aid the needed risk management decision 

—— techniques such as spatial and temporal analogues, e.g. analysis of scenarios that assess current vulnerabilities to past events (e.g., applying historical strong TC scenario to a modern city and estimating impacts) 

c) While socio-economic data are collected in many countries, permitting estimation of impacts, there are few ground-truth data available. We need to explore the use of social media to construct or validate impact models. 

d) Most datasets of population, socio-economic status, employment etc are static, based on infrequent censuses. However recent work has begun to explore combining such data to create time-dependent information. We need to further develop this capability to represent the dynamic nature of vulnerability. 

e) We need to identify and characterize vulnerabilities across populations for the selected hazards, including vulnerable populations who have not or only rarely exposed, e.g., Somalia TC, Myanmar TC 

f) We need to understand the reasons for counter-intuitive responses to risk warnings, and in particular to understand trade-offs in risks (e.g., fishermen going hungry vs. risking death when there is a high wave warning)